General Environmental Heterogeneity As the Explanation Of

bioRxiv preprint doi: https://doi.org/10.1101/141317; this version posted May 24, 2017. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

1 General environmental heterogeneity as the

2 explanation of sexuality? Comparative

3 study shows that ancient asexual taxa are

4 associated with both biotically and

5 abiotically homogeneous environments

7 Authors

8 Jan Toman*, Jaroslav Flegr

9 J. Toman, J. Flegr: Laboratory of Evolutionary Biology, Department of Philosophy and History 10 of Sciences, Faculty of Science, Charles University in Prague, Vinicna 7, 128 00 Prague 2, 11 Czech Republic

12 J. Toman: Tel.: +(420)221951821. Email: [email protected]

13 J. Flegr: Tel.: +(420)221951821. Email: [email protected]

14 * Corresponding author

15 Abstract

16 Ecological theories of sexual reproduction assume that sexuality is advantageous in certain 17 conditions, for example, in biotically or abiotically more heterogeneous environments. Such 18 theories thus could be tested by comparative studies. However, the published results of these 19 studies are rather unconvincing. Here we present the results of a new comparative study based 20 solely on the ancient asexual clades. The association with biotically or abiotically homogeneous 21 environments in these asexual clades was compared with the same association in their sister, or 22 closely related, sexual clades. Using the conservative definition of ancient asexuals (i.e. age > 1 23 million years), we found six pairs for which relevant ecological data are available. The difference 24 between the homogeneity type of environment associated with the sexual and asexual species 25 was then compared in an exact binomial test. Based on available literature, the results showed 26 that the vast majority of ancient asexual clades tend to be associated with biotically or 27 abiotically, biotically, and abiotically more homogeneous environments than their sexual 28 controls. In the exploratory part of the study, we found that the ancient asexuals often have 29 durable resting stages, enabling life in subjectively homogeneous environments, live in the 30 absence of intense biotic interactions, and are very often sedentary, inhabiting benthos and soil. 31 The consequences of these findings for the ecological theories of sexual reproduction are 32 discussed.

33 KEYWORDS: ancient asexuals – asexual reproduction – Frozen evolution theory – habitat 34 heterogeneity – sexual reproduction

35 Introduction

36 Paradox of sexual reproduction

37 The term “sexual” may be assigned to a broad array of processes (see e.g. Redfield, 2001; 38 Birky, 2009; Gorelick & Carpinone, 2009; Schön et al., 2009a). On the other hand, sex in its 39 strict sense (i.e. amphimixis) is defined as the alternation of meiosis and syngamy (Butlin et al., 40 1998a). 41 Sexual reproduction (sensu amphimixis) is one of the most enigmatic phenomena in 42 evolutionary biology, especially if its overwhelming predominance in eukaryotes is taken into 43 account (see e.g. Williams, 1975; Maynard Smith, 1978; Bell, 1982; Meirmans & Strand, 2010). 44 The reason is that it brings many obvious disadvantages in comparison to asexual reproduction— 45 the well-known two-fold cost of sex being only the first and most obvious one (see e.g. Lehtonen 46 et al., 2012). Thus, an asexual lineage invading the sexual population would be expected to 47 quickly drive its sexual conspecifics extinct (Maynard Smith, 1978). None of these 48 disadvantages apply to all sexual species because of the highly variable nature of their 49 reproduction. But, under many circumstances, the disadvantages apply profoundly (Lehtonen et 50 al., 2012). Thus, sexual reproduction remains an enigma that calls for explanation.

51 Theories of sexual reproduction

52 Many main concepts and their countless variants were proposed to explain the paradox of 53 sexual reproduction (reviewed e.g. in Williams, 1975; Maynard Smith, 1978; Bell, 1982, 1985; 54 Kondrashov, 1993; Meirmans & Strand, 2010). The genetic advantages of sexual reproduction as 55 the explanation for the maintenance of sexual reproduction are highlighted by theories like the 56 Weismann’s idea of sex generating variability, later delimited as the concept of Vicar of Bray 57 (Bell, 1982), Fisher-Muller’s accelerated evolution of sexual species (Muller, 1932; Fisher, 58 2003), breaking free of neighbouring deleterious mutations (Crow, 1970), reduction of the spread 59 of genomic parasites (Sterrer, 2002), advantage of diploidy (Lewis & Wolpert, 1979), repair of 60 DNA (Bernstein & Bernstein, 2013), restoration of epigenetic signals (Gorelick & Carpinone, 61 2009), eventually stochastic and deterministic variants of Muller’s ratchet hypothesis (Muller, 62 1964; Kondrashov, 1982). 63 Ecological theories of sexual reproduction, on the other hand, stress the assumption that 64 sex provides some ecological advantage to sexual species. Certain trends can be clearly found in

65 the geographic distribution of sexual reproduction, as was recently summarized by Hörandl 66 (2006, 2009) or Vrijenhoek & Parker (2009). As most advantages of sexual reproduction 67 postulated by genetic theories could be achieved by automixis (Gorelick & Carpinone, 2009; 68 Neiman & Schwander, 2011), the correct answer to the “greatest paradox of evolutionary 69 biology” probably lies in the group of ecological theories of sexual reproduction, or at least in 70 some form of theoretical synthesis that includes them (West et al., 1999).

71 Ecological theories of sexual reproduction and their predictions

72 Ecological theories such as Red Queen theory (Hamilton et al., 1990), evolutionary arm- 73 races hypothesis (Dawkins & Krebs, 1979), and the fast-sexual-response hypothesis of Maynard 74 Smith (1993) emphasize the sexually reproducing organisms’ advantage when interacting with 75 other organisms, which are able to dynamically react in a co-evolutionary manner. According to 76 these “biotic heterogeneity advantage” theories, sexual species should prosper in spatially and 77 temporally biotically heterogeneous environments, i.e. environments with many biotic 78 interactions from competitors, predators, and parasites – e.g. tropical rainforests, low-latitude 79 coral reefs, ancient lakes, climax communities, or generally species-rich ecosystems In such 80 environments, selective pressures affecting the offspring differ profoundly from those that 81 affected the parents because of the coadaptation (or rather counter-adaptation) between 82 interacting organisms. In the presence of such intensive biotic interactions, sexual species would 83 be especially favoured because they maintain high genetic polymorphism and could quickly react 84 to the countermoves of their evolutionary opponents by a simple change of allele frequencies in 85 the population. The speed, not the depth, of adaptation is essential in these environments 86 (Maynard Smith, 1993). 87 Another group of ecological theories of sexual reproduction sees the main advantage of 88 sexual reproduction in the higher fitness sexual individuals or species have in abiotically 89 heterogeneous environments. This group of “abiotic heterogeneity advantage” theories 90 comprises, for example, the lottery and Sisyphean genotypes hypothesis (Williams, 1975), elbow 91 room hypothesis (Maynard Smith, 1978), tangled bank hypothesis (Bell, 1982), hypothesis of 92 fluctuating selection (Smith, 1980), and hypothesis of reduced response to fluctuating selection 93 (Roughgarden, 1991). Such environments are abiotically very variable in space and/or time, i.e. 94 diverse, unpredictable, and with unequally distributed resources – e.g. temporary or ephemeral 95 habitats, exposed habitats, dynamically changing freshwater environments, coastal habitats or

96 more generally biomes of high latitudes and/or altitudes. The offspring usually inhabit an 97 environment different from that of their parents due to the dispersal in time and/or space. An 98 abiotic environment does not co-evolutionarily react to the evolutionary moves of its inhabitants, 99 allowing them to perfectly adapt to it under certain circumstances, e.g. under conditions of slow, 100 long-term changes. Under these circumstances, the asexual species might have an advantage 101 because, for example, they do not suffer from segregation and recombination loads (Crow, 102 1970). However, if the environment is very heterogeneous and unpredictable in time and/or 103 space, the sexual species have the same advantage as in the biotically heterogeneous 104 environment. 105 The heterogeneity of the environment, both biotic and abiotic, can be comprehended as 106 the sum of heterogeneity in space (in the sense of variability) and time (in the sense of instability, 107 especially when the change is unpredictable). The temporal and spatial aspects of heterogeneity, 108 even though differing substantially at first sight, could act remarkably similarly in terms of 109 favouring sexual species (Kondrashov, 1993; Otto & Lenormand, 2002; Neiman & Schwander, 110 2011). In principle, the most important factor is always whether the environment inhabited by the 111 offspring differs in its character (i.e. selective pressures) from the environment inhabited by their 112 parents. There are two conditions under which such a situation is fulfilled; the environment 113 either changes dynamically in time, or is very patchy and the offspring often live in different 114 microhabitats. 115 Both spatial and temporal heterogeneity could be the consequences of both biotic and 116 abiotic factors (Li & Reynolds, 1995). Thus, according to the major source of the environmental 117 heterogeneity, it is essentially possible to differentiate between “biotic” and “abiotic” ecological 118 theories of sexual reproduction, despite the fact that the biotic and abiotic parts of the 119 environmental heterogeneity are usually interconnected and that they influence and complement 120 each other in their effects on the advantage of sexual or asexual reproduction (Glesener & 121 Tilman, 1978). Abiotically homogeneous environments can (in the case of plentiful resources 122 and mild conditions), but do not have to (e.g. in the case of extreme environments) cause an 123 increase of the biotic aspect of environmental heterogeneity (e.g. by increasing the complexity of 124 the ecosystem) (Tokeshi, 1999). Abiotic heterogeneity could act both ways too. It might cause an 125 increase of biotic heterogeneity of the environment, or at least of the species diversity. On the

126 other hand, abiotically heterogeneous environments with conditions widely fluctuating to 127 extremes (e.g. the desiccating ponds) tend to be biotically homogeneous (Tokeshi, 1999). 128 The “biotic” and “abiotic” theories of sexual reproduction mentioned above have 129 different predictions regarding the character of the environment that will be advantageous for 130 sexual and asexual species. However, their predictions are not absolutely disparate—one could 131 easily devise examples of environments suitable for asexual species according to both groups of 132 theories, e.g. stable extreme environments. Similarly, the individual theories of sexual 133 reproduction are far from being disparate; they are usually interconnected in their basic 134 principles, they intermingle and complement each other (Meirmans & Strand, 2010). Many, if 135 not all “abiotic” theories can be reinterpreted “biotically”. It is therefore possible that the 136 differentiation of “biotic” and “abiotic” ecological theories is important in theory, but not 137 important in the real world, and that sexual organisms do have an advantage in environments that 138 are both biotically and abiotically relatively heterogeneous (i.e. overall heterogeneous 139 environments). 140 It was suggested by Williams (1975, pp. 145–146, 149–154, 169) and explicitly discussed 141 by Flegr (2013) that the lower ability of sexual species to fully adapt to transient environmental 142 changes brings them, paradoxically, a major advantage in randomly fluctuating environments, 143 i.e. in environments expressing large (biotic or abiotic) heterogeneity in time. Evolutionary 144 “plastic” (sensu Flegr) asexual species, sooner or later, adapt to transiently changed 145 environmental conditions and become extinct when the conditions return to normal. Sexual 146 species are not able to fully adapt to temporal environmental changes; they always retain some 147 genetic polymorphism that helps them escape extinction when the conditions quickly return to 148 normal. This hypothesis was supported by the results of certain experimental studies, e.g. long- 149 term patterns of fitness and genetic variability (Renaut et al., 2006) or dynamics of adaptation 150 (Colegrave et al., 2002; Kaltz & Bell, 2002) in sexually and asexually reproducing 151 Chlamydomonas. According to this (meta-)hypothesis, asexuals would prevail in stable or 152 predictively slowly changing, possibly extreme, environments of low temporal heterogeneity 153 (Flegr, 2013). Given the similarities between the effect of temporal and spatial heterogeneity 154 mentioned above, this notion can be readily extended to encompass both temporal and spatial 155 heterogeneity. Similar remarks were made, e.g. by Williams (1975, p. 153) and Roughgarden 156 (1991), while a combination of several aspects of heterogeneity was implicitly also proposed as

157 the explanation of the presence of sexual reproduction by Glesener & Tilman (1978) and some 158 interpreters of the Red Queen theory (e.g. Butlin et al., 1999) or the tangled bank hypothesis (e.g. 159 Bell, 1982).

160 Comparing the ecology of sexual and asexual groups

161 Most of the organisms that live on Earth, Archaea and Bacteria, are primarily asexual, 162 while most of the known species, eukaryotes, are primarily sexual (Speijer et al., 2015). 163 However, some eukaryotic lineages switched to secondary asexual reproduction (de Meeus, 164 Prugnolle & Agnew, 2007; Van Dijk, 2009; Speijer et al., 2015). It is therefore possible to 165 compare the environmental biotic and abiotic heterogeneity of such secondary asexual clades 166 with that of their sexual relatives. 167 Various studies aimed at testing and discriminating between individual ecological 168 theories of sexual reproduction on the basis of their predictions about the environmental 169 correlates of sexual and asexual lineages show largely inconclusive results. Often their aim was 170 to test particular theoretical concepts: lottery hypothesis and Sisyphean genotypes hypothesis 171 (Williams, 1975; Hörandl, 2009), elbow room hypothesis (Garcia & Toro, 1992; Koella, 1993), 172 Red Queen theory (Burt & Bell, 1987; Neiman & Koskella, 2009), fast-sexual-response 173 hypothesis (Becerra et al., 1999), hypothesis of optimal responsibility to fluctuating selection 174 (Griffiths & Butlin, 1995; Schön & Martens, 2004), hypothesis of prevention of loss of genetic 175 variability under fluctuating selection (Maynard Smith, 1993; Hörandl, 2009; Vrijenhoek & 176 Parker, 2009) or tangled bank hypothesis (Vrijenhoek, 1984; Burt & Bell, 1987; Griffiths & 177 Butlin, 1995; Domes et al., 2007; Maraun et al., 2012); or at least they were later interpreted as 178 such. The most extensive comparison not focused on testing one particular theoretical concept 179 was performed by Bell (1982) on multicellular animals (Metazoa). In his comprehensive study, 180 he concluded that the results mostly supported the tangled bank hypothesis. However, his study 181 included both old and young asexual taxa and might not fully appreciate the influence of 182 different life strategies on the character of the environment the species subjectively experience— 183 see the Discussion. Experiments aimed at discriminating the selective pressures of biotically (see 184 e.g. Fischer & Schmid-Hempel, 2005) or abiotically (see e.g. Becks & Agrawal, 2010) 185 heterogeneous and homogeneous environments were also performed, mostly pointing to the 186 conclusion that heterogeneous environments select higher rates of recombination or sexual

187 reproduction. However, particular mechanisms that favour higher levels of sex are hard to 188 determine in these cases that are, moreover, often based on facultatively sexual organisms.

189 Short-term and long-term asexual groups

190 The main problem of the comparative studies mentioned above may be the inclusion of 191 both old and young asexual taxa. Most secondary asexual groups probably are not evolutionarily 192 viable in the long term, as could be deduced from the distribution of asexual lineages on the “tree 193 of life.” With the exception of several ancient asexuals (AAs), they form only the terminal 194 twigs—species and genera (Butlin, 2002). This pattern could be the consequence of the 195 disadvantage that secondary asexual species have in species selection (Nunney, 1989). In the 196 large majority of cases, young asexuals do not speciate, and when the conditions suitable for 197 their opportunistic transition to asexuality change, they become extinct. Another suggested 198 reason for this macroevolutionary failure of asexual species is their failure in the interspecific 199 sorting in fluctuating environments—asexual species are able to adapt to temporarily changed 200 conditions more easily, but they fail to readapt quickly enough because of the lack of genetic 201 polymorphism, and therefore become extinct when the conditions return back to normal 202 (Williams, 1975; Flegr, 2013). At least some young asexual lineages could, in fact, consist of 203 short-lived clones continuously cleaved from maternal sexual population (Janko et al., 2008; 204 Vrijenhoek & Parker, 2009). Alternatively, they could be sustained by an occasional 205 hybridisation with related sexual lineages (Turgeon & Hebert, 1994; van Raay & Crease, 1995; 206 Butlin et al., 1998b) or an infrequent transfer of genetic material from “host species” in 207 hybridogenetic and gynogenetic lineages (Mantovani et al., 2001; Bogart et al., 2007). In sum, 208 young asexuals do not have to exhibit the properties that would allow them to survive in the long 209 term and, contrary to the mainstream view, they could in fact bring a significant noise into the 210 studies of long-term maintenance of sexual (and secondary asexual) reproduction. In other 211 words, the existence of most asexual lineages might be a consequence of certain ecological 212 opportunism and the reasons for their survival might be, in contrast to the AA lineages, different 213 from case to case. We speculate that this may be the main reason for the inconclusiveness of 214 most comparative studies mentioned above—they focused only on short-term asexual clones or 215 did not take into account the age of different clones. 216 The phenomenon of asexual “terminal twigs contra ancient asexuals” is still somewhat 217 controversial and its real existence is being discussed (see e.g. Schön et al., 1996; Neiman et al.,

218 2009; Schwander & Crespi, 2009; Janko et al., 2011). It was even suggested that the cumulative 219 frequency distribution of ages of asexual lineages is “fairly regular” and nearly linear on the 220 logarithmically transformed scale, i.e. it shows no obvious break between the young and old 221 asexuals (Neiman et al., 2009). As a consequence, AAs need not be substantially different from 222 other asexuals and their delimitation could be more or less arbitrary. On the other hand, Janko et 223 al. (2011) strongly questions the methodology of Neiman et al. (2009), especially the choice of 224 linear regression on logarithmic scale and its interpretation. They argue that the data distribution 225 presented by Neiman et al. (2009) is unlikely to provide any strong conclusions regarding the 226 universality of processes affecting clonal life spans—in fact, it fits even better to other 227 distributions, including a multidiffonential distribution, which is characteristic of distributions 228 generated by a mix of different underlying processes. Moreover, using their analysis of clonal 229 genetic variability, Janko et al. (2011) identified a growing selective pressure against asexual 230 lineages based on their age, corroborating the hypothesis of the long-term disadvantage of 231 asexuality. Regardless of these discussions, it is obvious that out of all the secondary asexual 232 clades only the AAs have been able to survive or even diversify in an asexual state for millions 233 of years (Judson & Normark, 1996; Normark et al., 2003; Neiman et al., 2009; Schurko et al., 234 2009; Schwander & Crespi, 2009). This is the main reason that our study is based exclusively on 235 AAs as they already proved to be evolutionarily viable in the long term. Contrastingly, young 236 opportunistic clonal lineages still face a serious extinction risk during the first pronounced 237 change of conditions. 238 However, it is worth mentioning that the focus on AAs puts forward another serious 239 difficulty: these clades were separated from their sister sexual lineages a long time ago (at least 1 240 million years ago, see Materials and Methods) and both sexual and asexual lineages thus 241 underwent considerable time periods of independent evolution. Therefore, both lineages 242 independently acquired numerous adaptations that distinguished them but need not be related to 243 the mode of their reproduction. Singular case studies comparing AAs and their sexual sister 244 lineage thus are not expected to have a strong predictive value in the long-term maintenance of 245 asexual reproduction. On the other hand, an extensive comparative study would enable the 246 comparison of several such pairs of AA and sexual control and reveal possible common 247 adaptations of AAs related to their long-term survival in an asexual state.

248 Aims of the study

249 The main aim of the current study is to test whether AAs more often inhabit 1) generally 250 less heterogeneous environments, 2) less biotically heterogeneous environments, or 3) less 251 abiotically heterogeneous environments. To this end, we used paired exact tests to compare the 252 ecological demands of sexual species and AA species within all six unrelated clades of 253 eukaryotic organisms in which the presence of differences in associated environmental 254 heterogeneity can be recognised on the basis of available published data. Namely, we tested the 255 hypothesis that, within such pairs, the asexual species inhabit more homogenous environment 256 more often than the sexual species. In the exploratory part of the study, we searched for 257 particular environmental properties and organismal adaptations that are common among the AA 258 members of the pairs.

259 Materials and Methods

260 Identification of ancient asexuals and their sexual controls

261 Ancient asexual groups 262 The definition of the “ancient asexual group” is rather vague. Some researchers consider 263 a lineage to be AA if it reproduces obligately asexually for at least 50 000 generations or 0.5 264 million years (Law & Crespi, 2002); some prefer one million generations (Schwander et al., 265 2011), yet others just speak about “millions of years” (Judson & Normark, 1996; Normark et al., 266 2003). It was even suggested that AAs are not substantially different from other asexuals and 267 their delimitation is more or less arbitrary (Neiman et al., 2009). It is not the aim of this study to 268 argue for the substantial difference of AAs from other asexuals or against it. We focus only on 269 groups that were proven to survive in an asexual state for a considerable amount of time. Thus, 270 regardless of the discussion on the fundamental distinction of young and old asexual taxa, which 271 was briefly presented in the Introduction, in the current study we defined AAs conservatively as 272 those secondary asexual eukaryotic lineages that reproduce obligately asexually with a great deal 273 of certainty for at least one million years. 274 The condition of obligate asexuality is at least equally as important as the age of the 275 studied asexual groups in the context of this study, as many seemingly asexual lineages might 276 experience rare sexual events or other forms of genetic exchange. These options constitute a

277 greater problem in short-term asexuals than in AAs, which are spatially and temporally isolated 278 from their sexual sister lineages. Nevertheless, their obligate asexuality was discussed, studied, 279 and verified thoroughly (see Table S1). 280 At the beginning, we identified well supported AA groups with the help of literary 281 sources. We started with published secondary literature such as Judson & Normark (1996), 282 Normark et al. (2003), Neiman et al. (2009), Schurko et al. (2009), Schwander & Crespi (2009) 283 and Speijer et al. (2015), investigated cited primary literature and other novel primary literal 284 sources concerning putative AA groups. We also investigated other possible AAs proposed in the 285 primary literature and some lineages traditionally believed to be long term asexual. The evidence 286 for confirmation or rejection of putative AAs included organismal, life-history, palaeontological, 287 biogeographical, molecular, individual genetic, and population genetic data and also other 288 indices of ancient asexuality proposed in the AA literature listed above. The list of supported and 289 contested AA candidates, as well as reasons for our decision, are summarized in Table S1. The 290 confirmed AA groups were eight: bdelloid rotifers (Bdelloidea), darwinulid ostracods 291 (Darwinulidae), several lineages of oribatid mites (Oribatidae), several lineages of mites from the 292 suborder Endeostigmata and order Trombidiformes, shoestring fern Vittaria appalachiana 293 (Farrar & Mickel), three species of stick insects from the genus Timema, and several lineages of 294 bivalve genus Lasaea. Only these groups were included in our comparative study.

295 Sexual controls 296 In the next step, we identified ecologically comparable sexual sister lineages for the eight 297 AA groups by using literary sources. In those individual cases in which the phylogenetic 298 relations between the sexual and asexual lineages were not entirely clear, we used the closest 299 possible comparable clades—clades proven to be closely related and comparable in terms of 300 their ecology (see Table S2). Three of the AA groups were monophyletic (Bdelloidea, 301 Darwinulidae, and Vittaria). The remaining AA groups were polyphyletic, i.e. they included 302 several related monophyletic asexual sub-lineages with interstitial sexual lineages. In each case 303 of a polyphyletic AA taxon, these closely related AA lineages were taken together and treated as 304 single unit in the analysis. In the opposite case, i.e. if the monophyletic lineages of these 305 polyphyletic taxa were incorporated into the analysis as single units, the risk would arise that the 306 observations would not be independent and the results of statistical analysis could be biased 307 because of the effect of pseudoreplications. Thus, in these cases, we compared every individual

308 AA lineage with its sister sexual lineage in the monophyletic sub-taxa of the polyphyletic AA 309 groups and based our conclusions on the prevailing trend (i.e. over 50 % of the cases; however, 310 all actual trends were much more convincing, see Table S3) in the polyphyletic groups. 311 Unfortunately, with the exception of Timema, the internal phylogenetic relationships of the 312 studied polyphyletic AA groups were more or less unclear. Where possible, we proceeded using 313 the most probable relationships (Bdelloidea, Darwinulidae, Oribatidae, Nematalycidae and 314 Proteonematalycidae, Grandjeanicidae and Oehserchestidae, see Table S2). In the cases with 315 several equally probable alternative phylogenetic relationships of AA and sexual lineages (both 316 in monophyletic /Vittaria/, and polyphyletic /Alicorhagia and Stigmalychus, Pomerantziidae, 317 Vittaria, Lasaea/ AA taxa), we compared AA lineages with alternative sexual controls to 318 determine the consistency of the trend in the association of AA lineages or sexual controls with 319 biotically and/or abiotically more heterogeneous environments (all trends were consistent over 320 all alternative sexual controls, see Table S3). 321 It could be argued that it is legitimate to consider particular asexual monophyletic 322 lineages in the polyphyletic AA group as independent observations when we inquire into the 323 reasons for maintaining sexual reproduction, not into the reasons for the transition to sexuality 324 (Thornhill & Fincher, 2013). However, we chose the more conservative approach because, as 325 noted above, the phylogenetic relationships of the individual AA and sexual lineages were often 326 not clear enough to form sisterly pairs of sexual and asexual lineages in the polyphyletic AA 327 groups with necessary certainty.

328 Determination of environmental heterogeneity

329 After the identification of the AA groups and their sexual controls, we focused on the 330 comparison of the character of their environments. Using relevant literary resources, we 331 collected and analysed data on the (biotically or abiotically more heterogeneous or 332 homogeneous) character of environments inhabited by the studied groups (see Table S3). Biotic 333 and abiotic environmental heterogeneity clearly have a non-trivial relationship to each other (see 334 the Discussion), but it is essentially possible to distinguish them. Environmental heterogeneity, 335 both biotic and abiotic, is an emergent property stemming from different factors and different 336 adaptations in various AAs and thus could not be quantified and rated on a single universal scale. 337 On the other hand, there is no obstacle of comparing AAs and their sexual controls based on this 338 general property of their environment.

339 Biotic heterogeneity 340 We define biotically highly heterogeneous environments as those in which selective 341 pressures affecting the offspring differ profoundly from those that previously affected their 342 parents because of the coadaptation (or rather counter-adaptation) of interacting organisms. 343 Thus, the most biotically heterogeneous environments are the habitats with a high degree of 344 competition, predation, and parasitism. Biotic heterogeneity has both temporal (the coadaptation 345 of interacting organisms) and spatial (e.g. the migration of the offspring to the areas with new 346 competitors, predators, and parasites) dimensions. Changes in the biotic heterogeneity are 347 essentially unpredictable, with the exception of some ecological cycles (e. g. host-predator or 348 host-parasite cycles). In the latter case, environments with unpredictable changes were 349 considered more biotically heterogeneous. 350 For example, the environment of organisms that live in tight association with other 351 organisms is biotically very heterogeneous. This applies especially to predators and parasites that 352 are forced to respond to the evolutionary counter-moves of their prey and hosts (Dawkins & 353 Krebs, 1979). The more specific relationship with prey and hosts they have, the stronger the 354 selective pressures of counter-adapting prey and hosts affect them (Dawkins & Krebs, 1979). 355 Therefore, it is expected that organisms that use non-specific predatory strategies, e.g., filtering 356 (especially if they filter both living organisms and dead organic matter), are under relatively 357 weak selective pressure from their prey. Their environment is consequently biotically relatively 358 homogenenous in this regard. On the other hand, the environment of organisms that are 359 themselves under strong selective pressure of predators and parasites is highly biotically 360 heterogeneous (Dawkins & Krebs, 1979). The environemnt of organisms that are not under 361 strong selective pressures of predators and parasites for various reasons is biotically more 362 homognenous in this regard (e.g. the environment of Darwinulidae, see Schön et al., 2009b or 363 Bruvo et al., 2011). Another important component of environmental biotic heterogeneity is 364 competition. The environments with complex ecosystems that are characterized by high a degree 365 of competition, predation and parasitism among their inhabitants (e.g. ancient lakes, see Martens, 366 1998; Martens & Schön, 2000; Schön & Martens, 2004) are highly biotically heterogeneous for 367 them. On the other hand, the environments with a low degree of competition, predation and 368 parasitism (especially extreme habitats, e.g. the environments of extremely high temperatures, 369 see Tobler, 2007, or, for photosynthetic organisms, poorly lit environments, see Farrar, 1978,

370 1998), are biotically very homogeneous for their inhabitants. The vast majority of environments 371 on Earth are thus somewhat biotically heterogeneous. In spite of that, we can find important 372 exceptions. This factor of environmental heterogeneity is considerably weakened especially in 373 extreme, ephemeral and marginal habitats that cannot sustain complex ecosystems because of 374 their extreme conditions, rapid unpredictable changes, low carrying capacity and/or insufficient 375 energy sources (see e.g. Bell, 1982). 376 An important factor that affects the biotic heterogeneity of environment the organisms 377 experience is the way of life practiced by aquatic organisms. On the average, lesser biotic 378 heterogeneity is experienced by benthic (or sedentary) organisms in comparison with planktonic 379 aquatic organisms. The reason is that the latter are subject to fast and effective transmission of 380 parasites and pathogens, especially viruses (see e.g. Suttle et al., 1990; Bratbak et al., 1993; 381 Fuhrman, 1999; Wommack & Colwell, 2000; Suttle, 2005, 2007), because of the character of 382 their environment—mixing of water masses lead to frequent encounters of various individuals 383 (Emiliani, 1993a, b). Crucial difference of the resulting risks for benthic and planktonic 384 organisms was pointed out by Emiliani (1982; 1993a). Emiliani (1993a) documented that benthic 385 representatives of Foraminifera have lower risk of extinction in comparison with planktonic 386 ones. The average length of existence of their benthic species was 20 million years, whereas 387 planktonic species lasted only about 7 million years. The most probable explanation of this 388 pattern is higher susceptibility of planktonic organisms to extinction caused by lethal parasitic, 389 especially viral, infections. This finding gave rise to the viral theory of background extinctions 390 (Emiliani 1993a, b). Evidence for the lower risks arising for benthic organisms from parasites 391 and pathogens was also supported by other ecological studies. For example, Filippini et al. 392 (2006) observed lower prevalence of individuals infected by viruses and consequent mortality 393 among benthic bacteria (~0,03 %) in comparison with bacteria from water column (~6 %). 394 Moreover, this pattern held despite much larger abundance of viruses in the sediment of studied 395 temperate lake. Fisher et al. (2003) and Bettarel et al. (2006) came to very similar conclusions on 396 the basis of studies of temperate oxbow lake and freshwater habitats in tropical Africa, 397 respectively (but see Danovaro et al., 2008 for somewhat contrasting results from marine benthic 398 sediment). Putting aside the limitations of parasite and pathogen transmission among benthic 399 organisms, the protection of benthic organisms against viruses might be further enhanced by the 400 adsorption of viruses into organic and inorganic particles of sediment and the aggregation of

401 benthic organisms (Filippini et al. 2006; Fisher et al., 2003). Whereas the extinction of whole 402 species due to viral infection (eventually infection by another parasite or pathogen, e.g. fungus) 403 is possible only under very limited conditions (see e.g. Buckwold, 1994, or de Castro & Bolker, 404 2005), local extinctions caused by pathogen or parasitic infections are probably quite common 405 (Emiliani, 1982, 1993a; de Castro & Bolker 2005). The planktonic way of life thus probably 406 considerably increases the selective pressures of parasites and pathogens and consequently the 407 biotic heterogeneity of the environment. 408 The argument for lower biotic heterogeneity of benthic (or sedentary) organisms is 409 possible to extend also to organisms that inhabit soil. From the viewpoint of the viral theory of 410 background extinctions (Emiliani, 1993a, b), soil represents an environment that considerably 411 impedes the spread of pathogens and parasites. Interactions with parasites and pathogens are very 412 limited both in intensity and frequency due to the tortuous, i.e., multidimensional, character of 413 soil matrix—it is best described as semi-discontinuous network of pores filled with air and/or 414 water, or water films surrounding solid particles (Lavelle & Spain, 2003; for more information 415 about the character of soil environment see also Wallwork, 1970; Coleman et al., 2004, or Paul, 416 2007). These features of soil matrix limit both the passive spread of parasites and pathogens and 417 the frequency of encounters among their transmitters and organisms in general. The tortuous 418 character of soil environments was stressed as a factor that limits the passive spread of viruses in 419 benthic sediment by Fisher et al. (2003), whereas Murphy & Tate (1996) emphasized its 420 limitations on the spread of bacteria. These observations are in agreement with Drake et al. 421 (1998), who observed a negative correlation between the concentration of viral particles and 422 sediment grain size. The sieving effect of the soil for organisms of various sizes is also 423 commented by Paul (2007). Moreover, the tortuous character of soil impedes also active 424 dispersion of organisms, e.g. when searching for prey (Elliott et al., 1980). Only a few larger 425 organisms are able to effectively move larger distances within soil or even create their own 426 habitats; movements of most organisms are locally constrained (Lavelle & Spain, 2003). Direct 427 encounters between organisms, even organisms of the same species, are thus relatively 428 infrequent. This leads, together with the limited spread of pheromones (Karasawa & Hijii, 2008), 429 for example, to frequent transitions to indirect fertilisation with the help of deposited 430 spermatophores (see e.g. Wallwork, 1970, or Lavelle & Spain, 2003). The pattern of spatial 431 autocorrelation of genetic lineages in soil communities, for example in rotifers (Robeson et al.,

432 2011), further supports the limited dispersal abilities of soil organisms. The genetic diversity of 433 various Bdelloidea lineages in soil is correlated only on small spatial scales (up to 54-133 m). 434 Operational taxonomic units identified by Robeson et al. (2011) almost did not overlap above 435 this distance. Habitats that distanced only tens to hundreds of meters were thus inhabited 436 overwhelmingly by separate genetic lineages. Moreover, rotifer communities differed to a certain 437 degree even in the smallest investigated distance of 16 cm (Robeson et al., 2011)1. 438 Taken together, the tortuous character of the soil affects all soil organisms at various 439 scales not only in terms of the reduced spread of parasites and pathogens, but also lower 440 frequencies of encounters with predators and competitors. This leads to an overall reduction of 441 biotic pressures in soil, which is further supported by the striking evolutionary stasis of many 442 lineages of soil inhabiting organisms (Pilato, 1979). Moreover, species richness and population 443 sizes, including parasites, predators and competitors, markedly decreases with the depth of soil 444 horizon (Lavelle & Spain, 2003; Paul, 2007). Deep soil horizons are therefore even more 445 abiotically homogeneous. The specific character of soil environment does not imply its general 446 spatial homogeneity. On the contrary, soil is often spatially heterogeneous, especially on a larger 447 scale (see e.g. Lavelle & Spain, 2003; Coleman, Crossley & Hendrix, 2004, or Paul, 2007). It is 448 the tortuous and multidimensional character of soil matrix that reduces biotic pressures affecting 449 its inhabitants and makes this environment biotically very homogeneous. 450 The biotic heterogeneity of the environment the organisms experience might be reduced 451 by the presence of durable resting stages. Organisms may get rid of parasites and pathogens, 452 survive unfavourable environmental conditions, or colonize new habitats with naïve parasites, 453 predators, pathogens and competitors in these stages (as do, for example, Bdelloidea—see 454 Wilson, 2011). The geographical trend of decreasing biotic heterogeneity with increasing latitude 455 might be expected on a global scale. Species diversity and ecosystem complexity decrease with 456 distance from equator (Tokeshi, 1999). These events are coupled with a decreasing intensity of 457 parasitization, abundance, prevalence and a relative diversity of parasites (Rohde, 1986; Rohde 458 & Heap, 1998). An analogous trend of decreasing biotic heterogeneity with increasing depth

1 High genetic diversity of Bdelloidea in gene cox1 is not very surprising in the light of severe DNA breaks that originate during anhydrobiosis, following repairs of these breaks and consequent intensive horizontal gene transfer (see e.g. Gladyshev et al., 2008).

459 might be expected in deeper parts of the water column for the same reasons (see e.g. Etter et al., 460 2005).

461 Abiotic heterogeneity 462 Abiotically highly heterogeneous environments are defined as those that are highly 463 variable regarding changes of abiotic factors. They are diverse, unstable, and have unequally 464 distributed resources. Again, the abiotic heterogeneity of the environment has both spatial (in the 465 sense of variability) and temporal (in the sense of instability) dimensions. The offspring thus 466 usually inhabit an environment different from that of their parents due to their dispersal in time 467 and/or space. Changes in the abiotic environment could be predictable (e.g. cyclical) or 468 unpredictable, and their intensity and frequency vary on different timescales. We are interested 469 in ecological timescales in this study so we consider short-term unpredictably changing 470 environments as the most abiotically heterogeneous. 471 Temporally and spatially highly changeable ephemeral and marginal habitats are 472 especially abiotically heterogeneous environments (see e.g. Pejler, 1995). However, most of the 473 surface terrestrial habitats are considerably abiotically heterogeneous. On the contrary, sheltered 474 habitats such as caves, ground water reservoirs or soil environment are greatly abiotically 475 homogenenous. Such environments protect their inhabitants from solar radiation and buffer 476 short-term fluctuations in outer environment (e.g. changes of temperature and humidity), 477 protecting their inhabitants from the direct impacts of such changes (Wallwork, 1970; Farrar, 478 1978, 1990, 1998; Krivolutsky & Druk, 1986; Lavelle & Spain, 2003; Devetter & Scholl, 2014). 479 Most changes in soil matrix are much slower in comparison with surface habitats (Lavelle & 480 Spain, 2003). The abiotic homogeneity of soil environment further increases with the depth of 481 the soil horizon. For example, there is a specific depth of soil horizon in each geographical 482 region under which the temperature is perienally stable, depending on its latitude, altitude and 483 other climatic factors (Wallwork, 1970; Lavelle & Spain, 2003; Coleman et al., 2004; Paul, 484 2007). The buffering effect of soil on moisture fluctuations also increases with depth (see e.g. 485 Quesada et al., 2004). Moreover, soils of certain biomes (especially forest soils) are temporally 486 abiotically more homogeneous than soils of other biomes (see e.g. Siepel, 1994, 1996). 487 Regarding aquatic environments, freshwater habitats and coastal areas are the most 488 abiotically heterogeneous (Sheldon, 1996). The decrease of abiotic heterogeneity with increasing 489 depth is also expected—water masses buffer surface environmental changes in a similar way to

490 soil (see e.g. Etter et al., 2005). Certain extreme environments that are temporally stable (e.g. hot 491 springs or subsurface cavities) are also very abiotically homogeneous (Bell, 1982), but this does 492 not apply to all environments referred to as extreme. 493 In a similar way to the biotic heterogeneity of the environment, also the abiotic one might 494 be reduced in the populations of organisms producing durable resting stages. Such an adaptation 495 enables them to survive unfavourable fluctuations of an abiotic environment and promotes 496 colonization of new habitats (see e.g. Wilson, 2011). On the other hand, mobility probably does 497 not strongly affect the abiotic environmental heterogeneity that the organisms experience. 498 Mobile organisms might hypothetically experience more different abiotic conditions in their life, 499 but they can also easily stick with those most suitable for them. The geographical trend of 500 increasing abiotic heterogeneity with increasing latitude and altitude might be expected to occur 501 on global scale (Hörandl, 2006, 2009; Vrijenhoek & Parker, 2009). However, it is noteworthy 502 that such a trend might be countered by an opposite trend of the decreasing biotic heterogeneity 503 mentioned above in its effects on sexual and asexual species (and vice versa). All of the 504 expectations mentioned above need not apply absolutely, but may serve as useful leads in 505 judging the environmental heterogeneity of various organisms if their peculiarities are taken into 506 account. 507 In two AA groups (Lasaea, Timema) we were unable to identify any consistent 508 differences in the heterogeneity of the environments inhabited by their sexual and asexual 509 lineages. The probability that any two groups experience absolutely identical habitat 510 heterogeneity is negligible and the absence of difference must be just a result of lack of empirical 511 data. Lasaea and Timema were thus not included in the statistical analysis because the binomial 512 test can analyse only binary variables (see e.g. McDonald, 2014) and our current knowledge on 513 these groups’ ecology does not allow us to decide whether asexual or sexual species live in more 514 heterogeneous environments. The same applies for the abiotic heterogeneity of the environment 515 of Darwinulidae. Thus, the analysis was based on only six pairs of AAs and their sexual controls, 516 out of which six differed in the abiotic heterogeneity of their environment and five in the biotic 517 heterogeneity.

518 Statistics

519 Collected data were analysed using the R v. 3.1.2 software environment (R_Core_Team, 520 2014). We chose to evaluate data by an exact test, specifically a one-tailed binomial test, due to

521 the low number of compared sexual-asexual pairs (six pairs in total, six for biotic heterogeneity 522 and five for abiotic heterogeneity), the character of the data (pair, binary), and the type of the 523 tested hypothesis. By this means, we tested three hypotheses, namely that AA groups inhabit 524 predominantly (1) biotically or abiotically, (2) biotically, and (3) abiotically more homogeneous 525 environments than their sister or closely related ecologically comparable sexual groups.

526 Results

527 Ancient asexuals

528 Reviewing relevant literary resources, including several recent studies, we concluded that 529 at least eight of the putative AA groups do fulfil our strict criteria of ancient asexuality: bdelloid 530 rotifers (Bdelloidea), darwinulid ostracods (Darwinulidae), some mite lineages of the Oribatidae, 531 Endeostigmata, and Trombidiformes, shoestring fern Vittaria appalachiana, some lineages of 532 stick insects from the genus Timema, and some lineages of the bivalve genus Lasaea; see 533 Supplemental table S1. Their sister or closely related ecologically comparable sexual groups 534 were identified consequently with the help of relevant literature; see Table S2.

535 Comparative analysis of the habitat homogeneity of ancient asexuals and their

536 sexual controls

537 The comparison of the character of environments inhabited by the AAs and their sexual 538 controls showed that AAs inhabit biotically or abiotically (6 of 6, p = 0.016), biotically (6 of 6, p 539 = 0.016), and abiotically (5 of 5, p = 0.031) more homogeneous environments. All these results 540 are statistically significant. Details of the results are summarized in the Supplemental review of 541 AA ecology and Table S3.

542 Exploratory part of the study

543 The exploratory part of this study is based on data gathered in the course of the main 544 analytic study. We identified several properties and adaptations that are common to a 545 considerable number of studied AAs. The most notable are durable resting stages, life in benthos 546 and soil, and life in the absence of intense biotic interactions. These findings are discussed 547 below.

548 Discussion

549 In contrast with other comparative studies in the field, the presented one is (1) based 550 exclusively on the AA taxa. Moreover, (2) biotic and abiotic environmental heterogeneity has 551 been strictly distinguished. We conclude that all 6 of the 6 analysed AA groups inhabit biotically 552 more homogeneous environments and all 5 of the 5 ones inhabit abiotically more homogeneous 553 environments when compared with their sexual controls. Some of the studied AA groups are 554 found in environments of a predominantly low biotic homogeneity (mainly Darwinulidae) or a 555 predominantly low abiotic homogeneity (mainly Oribatidae), but most of them are associated 556 with environments substantially more homogeneous both biotically and abiotically (Bdelloidea, 557 non-Oribatidae mites, Vittaria). No AA group lives in an environment abiotically or biotically 558 more heterogeneous than its sexual control. In the cases excluded from the analysis (abiotic 559 heterogeneity in Darwinulidae and both biotic and abiotic heterogeneity in Timema and Lasaea), 560 it was not possible to distinguish whether the heterogeneity is lower in the AA group or in the 561 sexual control. More specifically, enough information could not be found to decide either way. 562 The associations with biotically and abiotically more homogeneous environments overlap 563 almost perfectly. Thus, the results of the comparative analysis clearly indicate that either the AA 564 groups tend to be associated with overall (both biotically and abiotically) homogeneous 565 environments, or that these two types of heterogeneity are so strongly correlated that it is 566 impossible to decide in favour of theories of sexual reproduction that stress the key role of biotic 567 or abiotic heterogeneity. In general, our results obtained on AAs support, but of course do not 568 prove, the hypotheses that consider both biotic and abiotic heterogeneities acting as one factor in 569 their effect on organisms (Williams, 1975, pp. 145–146, 149–154, 169; Roughgarden, 1991; 570 Flegr, 2010, 2013). 571 Despite the widespread apprehension that the long independent evolution of AAs and 572 their sexual controls would hamper any ecological comparative analysis of the type presented 573 here (leading to the preference of young asexual lineages, see Introduction), we found that both 574 groups usually inhabit quite similar and considerably homogeneous environments. Contrary to 575 the initial expectation, this can, in fact, complicate analyses aimed at evaluating the differences 576 of the environments of AAs and their sexual controls in the opposite way (as was the case of 577 Timema and Lasaea, see Table S3). On the other hand, their common ancestor’s association with 578 the homogeneous environments could have been a preadaptation to the successful and long-term

579 transfer to asexual reproduction in the AAs. This tendency is obvious especially in 580 Darwinuloidea-Cypridoidea, but it can also be seen in Bdelloidea-Monogononta, Oribatidae, and 581 Endeostigmata (see Table S3). 582 It is interesting in this regard that a lot of contested AAs (see Table S1) also inhabit 583 considerably homogeneous environments—e.g. arbuscular mycorrhizal fungi of the order 584 Glomales (Croll & Sanders, 2009), tardigrades (Pilato, 1979; Mobjerg et al., 2011), nematode 585 genus Meloidogyne (Castagnonesereno et al., 1993), ostracods Heterocypris incongruens 586 (Ramdohr) and Eucypris virens (Jurine) (Butlin et al., 1998b; Martens, 1998), bristle fern 587 Trichomanes intricatum (Farrar) (Farrar, 1992), basidiomycete fungal families Lepiotaceae and 588 Tricholomataceae (Currie et al., 1999a, b), ambrosia fungi Ophiostomatales (Farrell et al., 2001), 589 or brine shrimp „Artemia parthenogenetica“ (Bowen & Sterling) (Vanhaecke et al., 1984)—and 590 their adaptations are similar to those of the AAs included in this study (see below).

591 What environmental properties and organismal adaptations are associated with AA

592 taxa?

593 The exploratory part of this study is based on data gathered in the course of the analytic 594 part of the study. Besides the tendency to inhabit biotically and abiotically homogeneous 595 environments, we discovered several properties and adaptations that are common to a 596 considerable number of studied AAs, do occur in AAs strikingly more often and could be the 597 particular adaptations enabling their long-term survival in the environments mentioned above. 598 Universally distributed adaptations potentially connected to the mode of reproduction were not 599 expected to be found in our sample because of markedly different ecological strategies of the 600 studied AAs. In spite of that, we found characteristics that have broad distribution among studied 601 taxa:

602 Alternative exchange of genetic information 603 Alternative ways of exchange of genetic information could theoretically substitute sexual 604 reproduction and thus were repeatedly proposed as the key adaptation to asexuality (Butlin et al., 605 1998a; Gladyshev et al. 2008; Boschetti et al. 2011; Debortoli et al., 2016; Schwander, 2016). 606 However, we identified this factor only once in the AAs included in our study (i.e. in 1/8 cases), 607 namely in Bdelloidea that experience intensive horizontal gene transfer (Gladyshev et al. 2008; 608 Boschetti et al. 2011; Debortoli et al., 2016). Another mechanism of genetic exchange,

609 parasexuality (sensu Pontecorvo 1954), was proposed in some contested ancient asexuals— 610 Glomales (Croll & Sanders, 2009), Tricholomataceae and Lepiotaceae (Mikheyev et al., 2006) 611 and certain protists (Birky, 2009). However, considering only the well supported AAs, these 612 mechanisms have limited distribution.

613 Durable resting stages and subjectively homogeneous environment 614 There is one particular, mostly unnoticed adaptation that could be essentially important 615 for the understanding of the links between the heterogeneity of environment and the mode of 616 reproduction. It is based on the insight that the character of the environment is “read” much 617 differently by its inhabitants than by the human observer. In case that a particular organism 618 reacts to the adverse change of environmental conditions by entrenching itself in the resting or 619 durable persistent stages (e.g. anabiosis) then, as a result, it de facto does not experience the 620 unfavourable conditions at all. Its objectively heterogeneous environment becomes subjectively 621 much more homogeneous. 622 This subjectivity addresses especially abiotic factors of environment, e.g. desiccation, 623 which is survived in the anabiotic stages by Bdelloidea (Pilato, 1979; Ricci, 2001), or freeze and 624 desiccation, which is survived in a state of torpor by Darwinulidae and some of their sexual 625 relatives (Carbonel et al., 1988). Similar durable stages could also be found in some contested 626 AAs, namely „Artemia parthenogenetica “(Vanhaecke et al., 1984) and tardigrades (Mobjerg et 627 al., 2011). Moreover, AA Lasaea is able to become mostly inactive and rests during the adverse 628 conditions for some time as well (Morton et al., 1957). On the other hand, at least in Bdelloidea, 629 the anhydrobiosis serves as the escape from biotic stresses too—especially parasites, both 630 directly (the individual gets rid of parasites during desiccation) and indirectly (by enabling the 631 escape from parasites in space and time) (Wilson, 2011). The distribution of durable resting 632 stages among well supported AAs looks rather scarce (3/8 cases). However, these 2-3 groups 633 comprise all studied AAs associated with significantly objectively abiotically heterogeneous 634 habitats. 635 Despite the undeniable importance of this phenomenon and the fact that it was repeatedly 636 highlighted, both historically (von Uexküll, 1982) and recently (Pilato, 1979), there is no general 637 awareness of the idea of difference between the subjective environment experienced by the 638 organism and the objective environment perceived by the human observer. This might be another 639 reason why most researchers did not come to unambiguous conclusions in their comparative

640 analyses of the ecology of sexual and asexual organisms. For example, a lot of “extreme” 641 environments are not necessarily abiotically very homogeneous, whereas some environments 642 that appear very abiotically heterogeneous on first sight (periodical ponds, dendrotelms etc.) 643 could be very subjectively homogeneous for local inhabitants (e.g. anhydrobiotic Bdelloidea). 644 After all, the “extremeness” of the environment itself depends on the adaptation of the observer, 645 including the presence or absence of the durable stages, and thus is highly subjective. 646 For a species, the ability to survive adverse conditions entrenched in durable persistent 647 stages could have crucial evolutionary consequences. It can be at least partially responsible for 648 the phenomenon of evolutionary stasis, great morphological uniformity, and conservatism of 649 numerous AA groups (see below) and indirectly responsible for the ability of successful and 650 long-term transition to asexuality. Such organisms are shielded against the selective pressures of 651 the environment. Their evolution is slow, but they are perfectly adapted to the conditions of their 652 habitat (Pilato, 1979).

653 Sedentary life and life in benthos 654 At least three well-supported AA groups (Bdelloidea, Darwinulidae, and Lasaea) are 655 exclusively benthic or sessile in contrast to their sexual relatives (Dole-Olivier et al., 2000; Ricci 656 & Balsamo, 2000; Ó Foighil, 1989). Some species of rotifer group Monogononta (sexual control 657 for Bdelloidea) (Pejler, 1995) and ostracod group Cypridoidea (sexual control for Darwinulidae) 658 (Martens et al., 2008) are planktonic; one of the two sexual lineages in genus Lasaea has 659 planktonic larvae (Ó Foighil, 1988). Life in benthos could theoretically hamper and reduce the 660 spread of parasites. Sessile organisms can resist hypothetical lethal epidemics postulated by the 661 viral theory of background extinction more efficiently (Emiliani, 1993a, b). 662 On the other hand, paleontological studies show that species with planktonic larvae have 663 decreased extinction rates in background extinction, particularly that species without planktonic 664 larvae go extinct more rapidly (Jablonski, 1986). In Lasaea, it seems that lineages with direct 665 development have a lower risk of extinction. These representatives also have a high chance of 666 speciation after migration into the new area (Ó Foighil & Eernisse, 1988), but the main reason is 667 probably that the direct developing lineages are surprisingly much better colonizers of distant 668 areas thanks to the rafting on vegetation (Ó Foighil, 1989). 669 In a similar way to resting stages, the distribution of benthic or sedentary lifestyle among 670 well supported AAs looks rather scarce on the first sight (3/8 cases). However, these 3 groups

671 comprise all studied AAs that are (at least partially) associated with aquatic habitats. It is 672 interesting that numerous contested aquatic AAs are also exclusively benthic: flatworm 673 Schmidtea polychroa (Schmidt) (Pongratz et al., 2003), New Zealand mudsnail Potamopyrgus 674 antipodarum (Gray) (Neiman et al., 2005), and ostracods Heterocypris incongruens (Ramdohr) 675 and Eucypris virens (Jurine) (Butlin et al., 1998b; Martens, 1998).

676 Life in the soil 677 Another adaptation widely distributed among AA groups is the inhabitancy of soil, 678 especially deeper parts of the soil horizon. This tendency can be seen mainly in the AA mites 679 from groups Oribatidae, Endeostigmata, and Trombidiformes (Karasawa & Hijii, 2008; Maraun 680 et al., 2009; Walter, 2009). Bdelloidea and Darwinuloidea tend to be associated with semi- 681 terrestrial habitats (Schön et al., 2009b). Moreover, AA Bdelloidea dominate among the soil 682 rotifers (Pejler, 1995). Most representatives of Darwinulidae inhabit soil (respectively interstitial) 683 too (Schön et al., 2009b). Taken together, 5/8 studied AA groups have numerous soil-inhabiting 684 representatives and a tendency to inhabit soil. Moreover, Bdelloidea dominate there over its 685 sexual control and AA mites have a stronger tendency to inhabit soil (Oribatidae) or its deep 686 horizons (Endeostigmata, Trombidiformes) in comparison with their sexual controls. 687 Living in soil could, in a similar way to life in benthos, reduce the capacity of parasites to 688 spread (sensu Emiliani, 1993a, b) Moreover, the tortuous character of the soil reduces any 689 interactions of soil organisms and thus negatively affect not only parasitization but also predation 690 and competition (Elliott et al., 1980; Murphy & Tate, 1996; Drake et al., 1998; Fisher et al., 691 2003; Lavelle & Spain, 2003; Paul, 2007). Besides, soil is an abiotically very stable environment 692 shielding its inhabitants from fluctuations in temperature and humidity, as well as from UV 693 radiation, and could be very favourable for asexuals also for this reason (Pilato, 1979; 694 Krivolutsky & Druk, 1986; Siepel, 1994). However, other explanations have also been proposed 695 for the asexuals’ association with soil habitats. Oribatidae could suffer less intense selective 696 pressures in the soil than in the arboreal environment where they have to respond to the co- 697 evolving lichens, their main food source (Maraun et al., 2009). Asexuality can be also more 698 advantageous in soil because of the difficulties with seeking out sexual partners, less effective 699 pheromone dispersal etc. (Karasawa & Hijii, 2008). Numerous contested AAs are soil 700 inhabitants too: Glomales (Croll & Sanders, 2009), tardigrades (Pilato, 1979; Jorgensen, 701 Mobjerg & Kristensen, 2007), and Meloidogyne (Castagnonesereno et al., 1993).

702 Thus, for the reasons mentioned above, life in soil habitats probably erases most of the 703 evolutionary advantages of sexuality. Selection pressures of biotic and abiotic environments are 704 weak in soil: The abiotic environment is relatively more stable and homogeneous; the 705 interactions with competitors, parasites, and predators are very limited both in intensity and 706 frequency due to the tortuous and multidimensional character of the environment (the shortest 707 way from point A to point B in soil is only rarely a straight line). The evolutionary consequences 708 of living in the soil are similar to those of having the capacity to form durable persistent stages 709 (see above). This is in accordance with the presence of noticeable evolutionary stasis of soil- 710 inhabiting groups described by Pilato (1979). It can also be seen in Bdelloidea (Ricci, 1987; 711 Poinar & Ricci, 1992) and Darwinulidae (Martens et al., 1998; Schön et al., 1998, 2009b). On 712 the other hand, the main factor causing evolutionary stasis in these two groups may be the 713 subjective homogeneity of the environment caused by the presence of dormant stages. However, 714 evolutionary stasis can be also found in other soil-inhabiting AAs: Oribatidae (Krivolutsky & 715 Druk, 1986; Norton, 1994; Heethoff et al., 2007) and other AA mites (Norton et al., 1993; 716 Walter, 2009; Walter et al., 2009); globally in 5/8 studied AA groups. Taking into account the 717 contested AA groups, it can be found in Glomales (Remy et al., 1994; Redecker, Kodner & 718 Graham, 2000) and tardigrades (Pilato, 1979; Jorgensen et al., 2007). All of these groups are 719 very uniform and have undergone almost no change in tens to hundreds of millions of years 720 (Pilato, 1979).

721 Absence of life strategies with intensive biotic interactions 722 It is noticeable that there are practically no typical predators and parasites among the AAs 723 we studied—this property is characteristic for all 8 studied groups. Remarkably often they feed 724 on dead organic matter or are autotrophic; parasites are almost absent, and in the case of a 725 predatory lifestyle, they are phytophagous or filtering (see Table S3). One possible explanation is 726 that they are unable to keep up in the co-evolutionary race with their sexual hosts or prey. Thus, 727 they can be successful in the long term, especially in the case of a predatory lifestyle, only if they 728 adopt (or are preadapted to) such nonspecific ecological strategies.

729 Succumbing to domestication and delegation of concern for its own benefit to another 730 biological entity 731 The tendency for asexual reproduction is particularly interesting in the contested AA 732 fungi domesticated by ants (Formicidae) and bark beetles (Scolytinae). The ant symbionts are 733 from the basidiomycete groups Tricholomataceae and Lepiotaceae (Mueller et al., 1998), 734 whereas bark beetles domesticate the ambrosia fungi of the ascomycete group Ophiostomatales 735 (Farrell et al., 2001). The association is particularly close in the ants. They care for the fungi 736 intensively, remove fungal predators and parasites, and the founding queen always carries 737 filamentous bacteria, which synthetize an antidote against the main fungal pathogen— 738 ascomycete Escovopsis (Currie et al., 1999a, b) —not to mention the stable temperature and 739 humidity in the nest. By doing so, they provide a very favourable, biotically and abiotically 740 stable environment. Moreover, there is some evidence that they prevent fungi from their already 741 minimal attempts at sexual reproduction. On the other hand, the situation may be more 742 complicated because some of these fungi create sexual structures predominantly in the presence 743 of ants (Mueller, 2002). This phenomenon provides an alternative view on some aspects of 744 human agriculture. A lot of crops raised by humans are asexual or self-pollinating too, which 745 probably facilitates their breeding but increases their susceptibility to parasites and pathogens 746 (Flegr, 2002). Life in association with another organism that takes care of the symbiont can also 747 be found in the contested AA group Glomales (Croll & Sanders, 2009) and various prokaryotic 748 and eukaryotic endosymbionts, see e.g. Douglas (2010). However, it has not been found in any 749 of the 8 well supported AA groups we studied, and its effect on the long-term maintenance of 750 asexual reproduction thus remains only speculative.

751 Conclusions

752 The analytical part of this study, i.e. the comparative analysis of the environment of AAs 753 and their sexual relatives, supported the hypothesis that AA groups are associated with overall 754 (biotically and abiotically) more homogeneous environments in comparison with their sister or 755 closely related ecologically comparable clades. It consequently supported the theoretical 756 concepts that postulate the essential advantage of sexual species in heterogeneous environments 757 and consider the (biotic and abiotic, temporal and spatial) heterogeneity of the environment 758 affecting the organisms to be one factor that can exhibit itself in many ways (Williams, 1975, pp.

759 145–146, 149–154, 169; Roughgarden, 1991; Flegr, 2010, 2013). Particular ecological 760 adaptations, from which durable resting stages, life in the absence of intense biotic interactions, 761 and the association with soil and benthic habitats are most notable, could be considered special 762 cases of the general AAs’ association with overall homogeneous environments. 763 Therefore, the general notion that proposed theories of sexual reproduction (see 764 Introduction) need not exclude each other, that the effects proposed by some or all of them might 765 intertwine and affect individuals and evolutionary lineages simultaneously, or that they even 766 may, ultimately, represent only different aspects of one more general explanation, seems to be 767 well supported. Moreover, overall environmental heterogeneity, regardless of its complicated 768 conceptualization and study, seems to be a suitable candidate for this hypothetical general 769 explanation. 770 Most putative AA lineages are still critically understudied. One way of elaborating the 771 foundations laid out by this study would be comparing the heterogeneity of environments in a 772 broader spectrum of AA lineages as soon as more lineages are discovered or confirmed (e.g. the 773 protist lineages proposed by Speijer et al., 2015). It would also be very desirable to investigate 774 the ecology of Lasaea, Timema, and Darwinulidae in greater detail. Additionally, it would be 775 appropriate to focus on the interaction of biotic and abiotic environmental heterogeneities and 776 their effect on organisms. According to Flegr (2008, 2010, 2013), sexual groups should exhibit 777 more pronounced evolutionary conservation of niches in comparison with asexuals—on the 778 whole, they are expected to stick closely around the phenotype of their common ancestor. This 779 hypothesis could be tested by comparing the variance of properties of individual species within 780 an AA and its related sexual clade. It would be also possible to test whether the sexual species 781 are able to survive under a wider range of conditions of the heterogeneous environment due to 782 their high genetic variability and hypothetical “elastic” reaction on selection, as was suggested 783 by Flegr (2008, 2010, 2013).

784 Acknowledgments

785 We are grateful to Radka Symonova, Karel Janko, Miloslav Devetter, Russell Shiel, Roy Norton, 786 Jaroslav Smrz, Jan Mourek, Lubomir Kovac, Vladimir Sustr, Peter Luptacik, Josef Stary, 787 Miroslav Kolarik, Lukas Kratochvil, Oldrich Fatka, Miroslav Kovarik, Adam Petrusek, Ivan 788 Cepicka, Vojtech Hampl and Marek Elias for their help with collecting information about

789 particular asexual taxa and useful insights into the biology of these species. We would also like 790 to thank Karel Kotrly, Vlasta Pachtova and Vojtech Zarsky for their help with acquiring some 791 hardly accessible literary sources and Eva Priplatova, Lenka Priplatova, Jinka Bousova, Julie 792 Nokola Novakova and Charlie Lotterman for the final revisions of our text.

793 Ethics approval and consent to participate

794 Not applicable

795 Consent for publication

796 Not applicable

797 Availability of data and materials

798 All data generated or analysed during this study are included in this published article and its 799 supporting information files.

800 Competing interests

801 The authors have no conflict of interests to declare.

802 Funding

803 The work was supported by Charles University in Prague under Grant project UNCE 204004.

804 Authors' contributions

805 JT gathered the data, prepared the figures and tables and was the greatest contributor in writing 806 the manuscript. JF contributed the analysis tools. Both JT and JF conceived and designed the 807 study, analysed the data and reviewed the drafts of the paper. All authors read and approved the 808 final manuscript.

809 Additional files

810 Additional file 1: E&E supporting information.pdf (TITLE: 3 Supplemental tables and 811 Supplemental review of AA ecology. DESRCRIPTION: Three Supplemental tables mentioned 812 in the manuscript and Supplemental review that corroborates the claims in Supplemental table 3 813 and explains them in more detail).

814 References

815 Becerra M, Brichette I, Garcia C. 1999. Short-term evolution of competition between genetically 816 homogeneous and heterogeneous populations of Drosophila melanogaster. Evolutionary 817 Ecology Research 1(5): 567-579. 818 Becks L, Agrawal AF. 2010. Higher rates of sex evolve in spatially heterogeneous environments. 819 Nature 468(7320): 89-93. 820 Bell G. 1982. The masterpiece of nature: the evolution and genetics of sexuality. London, UK: 821 Croom Helm. 822 Bell G. 1985. Two theories of sex and variation. Experientia 41(10): 1235-1245. 823 Bernstein H, Bernstein C. 2013. Evolutionary origin and adaptive function of meiosis. In: 824 Bernstein H, Bernstein C, eds. Meiosis. InTech, 41-75. Available at: 825 http://www.intechopen.com/books/meiosis/evolutionary-origin-and-adaptive-function-of- 826 meiosis (accessed 23 may 2017). DOI: 10.5772/56557 827 Bettarel Y, Bouvy M, Dumont C, Sime-Ngando T. 2006. Virus-bacterium interactions in water 828 and sediment of West African inland aquatic systems. Applied and Environmental 829 Microbiology 72(8): 5274-5282. 830 Birky CJ. 2009. Sex and evolution in eukaryotes. Encyclopedia of Life Support Systems 831 (EOLSS). Oxford, UK: Eolss Publishers. Available at: http://www.eolss.net/sample- 832 chapters/c03/e6-183-14-00.pdf (accessed 23 may 2017). 833 Bogart JP, Bi K, Fu JZ, Noble DWA, Niedzwiecki J. 2007. Unisexual salamanders (genus 834 Ambystoma) present a new reproductive mode for eukaryotes. Genome 50(2): 119-136. 835 Boschetti C, Pouchkina-Stantcheva N, Hoffmann P, Tunnacliffe A. 2011. Foreign genes and 836 novel hydrophilic protein genes participate in the desiccation response of the bdelloid 837 rotifer Adineta ricciae. Journal of Experimental Biology 214(1): 59-68. 838 Bratbak G, Egge JK, Heldal M. 1993. Viral mortality of the marine alga Emiliania huxleyi 839 (Haptophyceae) and termination of algal blooms. Marine Ecology Progress Series 93: 39- 840 48. 841 Bruvo R, Adolfsson S, Symonova R, Lamatsch D, Schön I, Jokela J, Butlin R, Muller S. 2011. 842 Few parasites, and no evidence for Wolbachia infections, in a freshwater ostracod 843 inhabiting temporary ponds. Biological Journal of the Linnean Society 102(1): 208-216. 844 Buckwold VE. 1994. Viral-induced extinctions unlikely. Nature 368(6471): 562.

845 Burt A, Bell G. 1987. Mammalian chiasma frequencies as a test of two theories of 846 recombination. Nature 326(6115): 803-805. 847 Butlin R. 2002. The costs and benefits of sex: new insights from old asexual lineages. Nature 848 Reviews Genetics 3(4): 311-317. 849 Butlin R, Schön I, Griffiths H. 1998a. Introduction to reproductive modes. In: Martens K, ed. 850 Sex and Parthenogenesis: Evolutionary Ecology of Reproductive Modes in Non-Marine 851 Ostracods. Leiden: Backhuys, 1-24. 852 Butlin R, Schön I, Martens K. 1998b. Asexual reproduction in nonmarine ostracods. Heredity 853 81(5): 473-480. 854 Butlin R, Schön I, Martens K. 1999. Origin, age and diversity of clones. Journal of Evolutionary 855 Biology 12(6): 1020-1022. 856 Carbonel P, Colin J, Danielopol D, Loffler H, Neustrueva I. 1988. Paleoecology of limnic 857 ostracodes: a review of some major topics. Palaeogeography Palaeoclimatology 858 Palaeoecology 62(1-4): 413-461. 859 Castagnonesereno P, Piotte C, Uijthof J, Abad P, Wajnberg E, Vanlerberghemasutti F, 860 Bongiovanni M, Dalmasso A. 1993. Phylogenetic relationships between amphimictic and 861 parthenogenetic nematodes of the genus Meloidogyne as inferred from repetitive DNA 862 analysis. Heredity 70(2): 195-204. 863 Colegrave N, Kaltz O, Bell G. 2002. The ecology and genetics of fitness in Chlamydomonas. 864 VIII. The dynamics of adaptation to novel environments after a single episode of sex. 865 Evolution 56(1): 14-21. 866 Coleman DA, Crossley DA, Hendrix PF. 2004. Fundamentals of Soil Ecology, Second edition. 867 USA: Elsevier Academic Press. 868 Croll D, Sanders I. 2009. Recombination in Glomus intraradices, a supposed ancient asexual 869 arbuscular mycorrhizal fungus. Bmc Evolutionary Biology 9(1). 870 Crow J. 1970. Genetic loads and the cost of natural selection. In: Kojima K, ed. Biomathematics. 871 Volume 1. Mathematical topics in population genetics. Berlin: Springer-Verlag, 128-177. 872 Currie C, Mueller U, Malloch D. 1999a. The agricultural pathology of ant fungus gardens. 873 Proceedings of the National Academy of Sciences of the United States of America 96(14): 874 7998-8002.

875 Currie C, Scott J, Summerbell R, Malloch D. 1999b. Fungus-growing ants use antibiotic- 876 producing bacteria to control garden parasites. Nature 398(6729): 701-704. 877 Danovaro R, Dell’Anno A, Corinaldesi C, Magagnini M, Noble R, Tamburini C, Weinbauer M. 878 2008. Major viral impact on the functioning of benthic deep-sea ecosystems. Nature 879 454(7208): 1084-1087. 880 Dawkins R, Krebs JR. 1979. Arms races between and within species. Proceedings of the Royal 881 Society B-Biological Sciences 205(1161): 489-511. 882 Debortoli N, Li X, Eyres I, Fontaneto D, Hespeels B, Tang CQ, Flot J, Van Doninck K. 2016. 883 Genetic exchange among bdelloid rotifers is more likely due to horizontal gene transfer 884 than to meiotic sex. Current Biology 26(6): 723-732. 885 de Castro F, Bolker B. 2005. Mechanisms of disease-induced extinction. Ecology Letters 8(1): 886 117-126. 887 de Meeus T, Prugnolle F, Agnew P. 2007. Asexual reproduction: genetics and evolutionary 888 aspects. Cellular and Molecular Life Sciences 64(11): 1355-1372. 889 Devetter M, Scholl K. 2014. Hydrobiont animals in floodplain soil: are they positively or 890 negatively affected by flooding? Soil Biology & Biochemistry 69: 393-397. 891 Dole-Olivier M, Galassi D, Marmonier P, Des Chatelliers M. 2000. The biology and ecology of 892 lotic microcrustaceans. Freshwater Biology 44(1): 63-91. 893 Domes K, Scheu S, Maraun M. 2007. Resources and sex: soil re-colonization by sexual and 894 parthenogenetic oribatid mites. Pedobiologia 51(1): 1-11. 895 Douglas A. 2010. The Symbiotic Habit. Princeton, NJ: Princeton University Press. 896 Drake LA, Choi KH, Haskell AE, Dobbs FC. 1998. Vertical profiles of virus-like particles and 897 bacteria in the water column and sediments of Chesapeake Bay, USA. Aquatic Microbial 898 Ecology 16(1): 17-25. 899 Elliott ET, Anderson RV, Coleman DC, Cole CV. 1980. Habitable pore space and microbial 900 trophic interactions. Oikos 35(3): 327-335. 901 Emiliani C. 1982. Extinctive evolution: extinctive and competitive evolution combine into a 902 unified model of evolution. Journal of Teoretical Biology 97(1): 13-33. 903 Emiliani C. 1993a. Extinction and viruses. Biosystems 31(2-3): 155-159. 904 Emiliani C. 1993b. Viral extinctions in deep-sea species. Nature 366(6452): 217-218.

905 Etter RJ, Rex MA, Chase MR, Quattro JM. 2005. Population differentiation decreases with depth 906 in deep-sea bivalves. Evolution 59(7): 1479-1491. 907 Farrar D. 1978. Problems in the identity and origin of the Appalachian Vittaria gametophyte, a 908 sporophyteless fern of the eastern United States. American Journal of Botany 65(1): 1-12. 909 Farrar D. 1990. Species and evolution in asexually reproducing independent fern gametophytes. 910 Systematic Botany 15(1): 98-111. 911 Farrar D. 1992. Trichomanes intricatum: the independent Trichomanes gametophyte in the 912 eastern United States. American Fern Journal 82(2): 68-74. 913 Farrar D. 1998. The tropical flora of rockhouse cliff formations in the eastern United States. 914 Journal of the Torrey Botanical Society 125(2): 91-108. 915 Farrell BD, Sequeira AS, O'Meara BC, Normark BB, Chung JH, Jordal BH. 2001. The evolution 916 of agriculture in beetles (Curculionidae: Scolytinae and Platypodinae). Evolution 55(10): 917 2011-2027. 918 Filippini M, Buesing N, Bettarel Y, Sime-Ngando T, Gessner MO. 2006. Infection paradox: high 919 abundance but low impact of freshwater benthic viruses. Applied and Environmental 920 Microbiology 72(7): 4893-4898. 921 Fischer O & Schmid-Hempel P. 2005. Selection by parasites may increase host recombination 922 frequency. Biology Letters 1(2): 193–195. 923 Fisher R. 2003. The Genetical Theory of Natural Selection: A Complete Variorum Edition. New 924 York, NY: Oxford University Press. 925 Fisher UR, Wieltschnig C, Kirschner AKT, Velimirov B. 2003. Does virus-induced lysis 926 contribute significantly to bacterial mortality in the oxygenated sediment layer of shallow 927 oxbow lakes? Applied and Environmental Microbiology 69(9): 5281-5289. 928 Flegr J. 2002. Was Lysenko (partly) right? Michurinist biology in the view of modern plant 929 physiology and genetics. Rivista Di Biologia-Biology Forum 95: 259-271. 930 Flegr J. 2008. Frozen evolution: Or, that's not the way it is, mr. Darwin - Farewell to selfish 931 gene. USA: Createspace Independent Pub. 932 Flegr J. 2010. Elastic, not plastic species: frozen plasticity theory and the origin of adaptive 933 evolution in sexually reproducing organisms. Biology Direct 5(1). 934 Flegr J. 2013. Microevolutionary, macroevolutionary, ecological and taxonomical implications 935 of punctuational theories of adaptive evolution. Biology Direct 8(1).

936 Fuhrman JA. 1999. Marine viruses and their biogeochemical and ecological effects. Nature 937 399(6736): 541-548 938 Garcia C, Toro M. 1992. Sib competition in Tribolium: a test of the elbow-room model. Heredity 939 68(6): 529-536. 940 Gladyshev E, Meselson M. 2008. Extreme resistance of bdelloid rotifers to ionizing radiation. 941 Proceedings of the National Academy of Sciences of the United States of America 942 105(13): 5139-5144. 943 Glesener R, Tilman D. 1978. Sexuality and the components of environmental uncertainty: clues 944 from geographic parthenogenesis in terrestrial animals. The American Naturalist 945 112(986): 659-673. 946 Gorelick R, Carpinone J. 2009. Origin and maintenance of sex: the evolutionary joys of self sex. 947 Biological Journal of the Linnean Society 98(4): 707-728. 948 Griffiths H, Butlin R. 1995. A timescale for sex versus parthenogenesis: evidence from subfossil 949 ostracods. Proceedings of the Royal Society of London Series B-Biological Sciences 950 260(1357): 65-71. 951 Hamilton W, Axelrod R, Tanese R. 1990. Sexual reproduction as an adaptation to resist parasites 952 (a review). Proceedings of the National Academy of Sciences of the United States of 953 America 87(9): 3566-3573. 954 Heethoff M, Domes K, Laumann M, Maraun M, Norton RA, Scheu S. 2007. High genetic 955 divergences indicate ancient separation of parthenogenetic lineages of the oribatid mite 956 Platynothrus peltifer (Acari, Oribatida). Journal of Evolutionary Biology 20(1): 392-402. 957 Hörandl E. 2006. The complex causality of geographical parthenogenesis. New Phytologist 958 171(3): 525-538. 959 Hörandl E. 2009. Geographical parthenogenesis: opportunities for asexuality. In: Schön I, 960 Martens K, van Dijk P, eds. Lost Sex: The Evolutionary Biology of Parthenogenesis. 961 Dordrecht: Springer Science+Business Media B.V., 161-186. 962 Jablonski D. 1986. Background and mass extinctions: the alternation of macroevolutionary 963 regimes. Science 231(4734): 129-133. 964 Janko K, Drozd P, Eisner J. 2011. Do clones degenerate over time? Explaining the genetic 965 variability of asexuals through population genetic models. Biology Direct 6(1).

966 Janko K, Drozd P, Flegr J, Pannell J. 2008. Clonal turnover versus clonal decay: a null model for 967 observed patterns of asexual longevity, diversity and distribution. Evolution 62(5): 1264- 968 1270. 969 Jorgensen A, Mobjerg N, Kristensen R. 2007. Molecular study of the tardigrade Echiniscus 970 testudo (Echiniscidae) reveals low DNA sequence diversity over a large geographical 971 area. Journal of Limnology 66(1s): 77-83. 972 Judson OP, Normark BB. 1996. Ancient asexual scandals. Trends in Ecology & Evolution 11(2): 973 41-46. 974 Kaltz O, Bell G. 2002. The ecology and genetics of fitness in Chlamydomonas. XII. Repeated 975 sexual episodes increase rates of adaptation to novel environments. Evolution 56(9): 976 1743-1753. 977 Karasawa S, Hijii N. 2008. Vertical stratification of oribatid (Acari: Oribatida) communities in 978 relation to their morphological and life-history traits and tree structures in a subtropical 979 forest in Japan. Ecological Research 23(1): 57-69. 980 Koella J. 1993. Ecological correlates of chiasma frequency and recombination index of plants. 981 Biological Journal of the Linnean Society 48(3): 227-238. 982 Kondrashov A. 1982. Selection against harmful mutations in large sexual and asexual 983 populations. Genetical Research 40(03): 325-332. 984 Kondrashov A. 1993. Classification of hypotheses on the advantage of amphimixis. Journal of 985 Heredity 84(5): 372-387. 986 Krivolutsky D, Druk A. 1986. Fossil oribatid mites. Annual Review of Entomology 31(1): 533- 987 545. 988 Lavelle P, Spain AV. 2003. Soil Ecology. New York, USA: Kluwer Academic Publishers. 989 Law J, Crespi B. 2002. Recent and ancient asexuality in Timema walkingsticks. Evolution 56(8): 990 1711-1717. 991 Lehtonen J, Jennions M, Kokko H. 2012. The many costs of sex. Trends in Ecology & Evolution 992 27(3): 172-178. 993 Lewis J, Wolpert L. 1979. Diploidy, evolution and sex. Journal of Theoretical Biology 78(3): 994 425-438. 995 Li H, Reynolds J. 1995. On definition and quantification of heterogeneity. Oikos 73(2): 280-284.

996 Mantovani B, Passamonti M, Scali V. 2001. The mitochondrial cytochrome oxidase II gene in 997 Bacillus stick insects: ancestry of hybrids, androgenesis, and phylogenetic relationships. 998 Molecular Phylogenetics and Evolution 19(1): 157-163. 999 Maraun M, Erdmann G, Schulz G, Norton RA, Scheu S, Domes K. 2009. Multiple convergent 1000 evolution of arboreal life in oribatid mites indicates the primacy of ecology. Proceedings 1001 of the Royal Society B-Biological Sciences 276(1671): 3219-3227. 1002 Maraun M, Norton RA, Ehnes RB, Scheu S, Erdmann G. 2012. Positive correlation between 1003 density and parthenogenetic reproduction in oribatid mites (Acari) supports the structured 1004 resource theory of sexual reproduction. Evolutionary Ecology Research 14(3): 311-323. 1005 Martens K. 1998. Sex and ostracods: a new synthesis. In: Martens K, ed. Sex and 1006 Parthenogenesis: Evolutionary Ecology of Reproductive Modes in Non-Marine Ostracds. 1007 Leiden: Backhuys, 295-321. 1008 Martens K, Horne D, Griffiths H. 1998. Age and diversity of non-marine ostracods. In: Martens 1009 K, ed. Sex and Parthenogenesis: Evolutionary Ecology of Reproductive Modes in Non- 1010 Marine Ostracds. Leiden: Backhuys, 37-55. 1011 Martens K, Schön I. 2000. The importance of habitat stability for the prevalence of sexual 1012 reproduction. In: Minoura K, ed. Lake Baikal: a Mirror in Time and Space for 1013 Understanding Global Change Processes. Yokohama Symposium 1998. Amsterdam: 1014 Elsevier, 324-330. 1015 Martens K, Schön I, Meisch C, Horne D. 2008. Global diversity of ostracods (Ostracoda, 1016 Crustacea) in freshwater. Hydrobiologia 595(1): 185-193. 1017 Maynard Smith J. 1978. The evolution of sex. Cambridge, UK: Cambridge University Press. 1018 Maynard Smith J. 1993. The theory of evolution. Cambridge, UK: Cambridge University Press. 1019 McDonald JH. 2014. Handbook of Biological Statistics (3rd ed.). Baltimore, USA: Sparky 1020 House Publishing. 1021 Meirmans S, Strand R. 2010. Why are there so many theories for sex, and what do we do with 1022 them? Journal of Heredity 101(1s): S3-S12. 1023 Mikheyev AS, Mueller UG, Abbot P. 2006. Cryptic sex and many-to-one colevolution in the 1024 fungus-growing ant symbiosis. Proceedings of the National Academy of Sciences of the 1025 United States of America 103(28): 10702-10706.

1026 Mobjerg N, Halberg K, Jorgensen A, Persson D, Bjorn M, Ramlov H, Kristensen R. 2011. 1027 Survival in extreme environments – on the current knowledge of adaptations in 1028 tardigrades. Acta Physiologica 202(3): 409-420. 1029 Morton JE, Boney AD, Corner EDS. 1957. The adaptations of Lasaea rubra (Montagu), a small 1030 intertidal lamellibranch. Journal of the Marine Biological Association of the United 1031 Kingdom 36(02): 383-405. 1032 Mueller UG. 2002. Ant versus fungus versus mutualism: ant–cultivar conflict and the 1033 deconstruction of the attine ant–fungus symbiosis. American Naturalist 160(4s): S67- 1034 S98. 1035 Mueller UG, Rehner SA, Schultz TR. 1998. The evolution of agriculture in ants. Science 1036 281(5385): 2034-2038. 1037 Muller H. 1932. Some genetic aspects of sex. The American Naturalist 66(703): 118-138. 1038 Muller H. 1964. The relation of recombination to mutational advance. Mutation Research 1(1): 1039 2-9. 1040 Murphy SL, Tate RL. 1996. Bacterial movement through soil. In: Stozsky G & Bollag JM, ed. 1041 Soil Biochemistry Vol. 9. New York: Marcel Dekker, 253-286. 1042 Neiman M, Jokela J, Lively C. 2005. Variation in asexual lineage age in Potamopyrgus 1043 antipodarum, a New Zealand snail. Evolution 59(9): 1945-1952. 1044 Neiman M, Koskella B. 2009. Sex and the Red queen. In: Schön I, Martens K, van Dijk P, eds. 1045 Lost sex: the Evolutionary Biology of Parthenogenesis. Dordrecht: Springer 1046 Science+Business Media B.V., 133-159. 1047 Neiman M, Meirmans S, Meirmans P, Schlichting C, Mousseau T. 2009. What can asexual 1048 lineage age tell us about the maintenance of sex? Annals of the New York Academy of 1049 Sciences 1168(1): 185-200. 1050 Neiman M, Schwander T. 2011. Using parthenogenetic lineages to identify advantages of sex. 1051 Evolutionary Biology 38(2): 115-123. 1052 Normark B, Judson O, Moran N. 2003. Genomic signatures of ancient asexual lineages. 1053 Biological Journal of the Linnean Society 79(1): 69-84. 1054 Norton R. 1994. Evolutionary aspects of oribatid mite life histories and consequences for the 1055 origin of the Astigmata. In: Houck M, ed. Mites. Ecological and Evolutionary Analyses 1056 of Life-History Patterns. New York: Chapman & Hall, 99-135.

1057 Norton R, Kethley J, Johnston D, O'Connor B. 1993. Phylogenetic perspectives on genetic 1058 systems and reproductive modes of mites. In: Wrensch D, Ebbert M, eds. Evolution and 1059 diversity of sex ratio in insects and mites. New York: Chapman and Hall, 8-99. 1060 Nunney L. 1989. The maintenance of sex by group selection. Evolution 43(2): 245-257. 1061 Otto S, Lenormand T. 2002. Resolving the paradox of sex and recombination. Nature Reviews 1062 Genetics 3(4): 252-261. 1063 Ó Foighil D. 1988. Random mating and planktotrophic larval development in the brooding 1064 hermaphroditic clam Lasaea australis (Lamarck, 1818). Veliger 31(3-4): 214-221. 1065 Ó Foighil D. 1989. Planktotrophic larval development is associated with a restricted geographic 1066 range in Lasaea, a genus of brooding, hermaphroditic bivalves. Marine Biology 103(3): 1067 349-358. 1068 Ó Foighil D, Eernisse D. 1988. Geographically widespread, non-hybridizing, sympatric strains of 1069 the hermaphroditic, brooding clam Lasaea in the northeastern Pacific Ocean. Biological 1070 Bulletin 175(2): 218-229. 1071 Paul EA. 2007. Soil Microbiology, Ecology and Biochemistry. USA: Academic Press. 1072 Pejler B. 1995. Relation to habitat in rotifers. Hydrobiologia 313(1): 267-278. 1073 Pilato G. 1979. Correlations between cryptobiosis and other biological characteristics in some 1074 soil animals. Italian Journal of Zoology 46(4): 319-332. 1075 Poinar GO, Ricci C. 1992. Bdelloid rotifers in Dominican amber: evidence for parthenogenetic 1076 continuity. Experientia 48(4): 408-410. 1077 Pongratz N, Storhas M, Carranza S, Michiels N. 2003. Phylogeography of competing sexual and 1078 parthenogenetic forms of a freshwater flatworm: patterns and explanations. Bmc 1079 Evolutionary Biology 3(1). 1080 Pontecorvo D. 1954. Mitotic recombination in the genetic systems of filamentous fungi. 1081 Caryologia 6: 192-200. 1082 Quesada CA, Miranda AC, Hodnett MG, Santos AJB, Miranda HS, Breyer LM. 2004. Seasonal 1083 and depth variation of soil moisture in a burned open savanna (campo sujo) in central 1084 Brazil. Ecological Applications 14(4s): 33-41. 1085 Redecker D, Kodner R, Graham L. 2000. Glomalean fungi from the Ordovician. Science 1086 289(5486): 1920-1921. 1087 Redfield R. 2001. Do bacteria have sex? Nature Reviews Genetics 2(8): 634-639.

1088 Remy W, Taylor T, Hass H, Kerp H. 1994. Four hundred-million-year-old vesicular arbuscular 1089 mycorrhizae. Proceedings of the National Academy of Sciences of the United States of 1090 America 91(25): 11841-11843. 1091 Renaut S, Replansky T, Heppleston A, Bell G. 2006. The ecology and genetics of fitness in 1092 Chlamydomonas. XIII. The fitness of long-term sexual and asexual populations in benign 1093 environments. Evolution 60(11): 2272-2279. 1094 Ricci C. 2001. Dormancy patterns in rotifers. Hydrobiologia 446(1): 1-11. 1095 Ricci C, Balsamo M. 2000. The biology and ecology of lotic rotifers and gastrotrichs. 1096 Freshwater Biology 44(1): 15-28. 1097 Ricci CN. 1987. Ecology of bdelloids: how to be successful. Hydrobiologia 147(1): 117-127. 1098 Robeson MS, King AJ, Freeman KR, Birky CW, Martin AP, Schmidt SK. 2011. Soil rotifer 1099 communities are extremely diverse globally but spatially autocorrelated locally. 1100 Proceedings of the National Academy of Sciences of the United States of America 1101 108(11): 4406-4410. 1102 Rohde K. 1986. Differences in species diversity of monogenea between the pacific and atlantic 1103 oceans. Hydrobiologia 137(1): 21-28. 1104 Rohde K, Heap M. 1998. Latitudinal differences in species and community richness and in 1105 community structure of metazoan endo- and ectoparasites of marine teleost fish. 1106 International Journal For Parasitology 28(3): 461-474. 1107 Roughgarden J. 1991. The evolution of sex. American Naturalist 138(4): 934-953. 1108 R_Core_Team. 2014. R: A language and environment for statistical computing, version 3.1.2. 1109 Vienna, Austria: R Foundation for Statistical Computing. Available at: https://www.r- 1110 project.org (accessed 23 may 2016). 1111 Schurko A, Neiman M, Logsdon J. 2009. Signs of sex: what we know and how we know it. 1112 Trends in Ecology & Evolution 24(4): 208-217. 1113 Schwander T, Crespi BJ. 2009. Twigs on the tree of life? Neutral and selective models for 1114 integrating macroevolutionary patterns with microevolutionary processes in the analysis 1115 of asexuality. Molecular Ecology 18(1): 28-42. 1116 Schwander T, Henry L, Crespi B. 2011. Molecular evidence for ancient asexuality in Timema 1117 stick insects. Current Biology 21(13): 1129-1134.

1118 Sheldon P. 1996. Plus ça change—a model for stasis and evolution in different environments. 1119 Palaeogeography, Palaeoclimatology, Palaeoecology 127(1-4): 209–227. 1120 Schön I, Butlin R, Griffiths H, Martens K. 1998. Slow molecular evolution in an ancient asexual 1121 ostracod. Proceedings of the Royal Society of London Series B-Biological Sciences 1122 265(1392): 235-242. 1123 Schön I, Martens K. 2004. Adaptive, pre-adaptive and non-adaptive components of radiations in 1124 ancient lakes: a review. Organisms Diversity & Evolution 4(3): 137-156. 1125 Schön I, Martens K, Rossi V. 1996. Ancient asexuals: scandal or artifact? Trends in Ecology and 1126 Evolution 11(7): 296-297. 1127 Schön I, Martens K, van Dijk P. 2009a. Lost Sex: The Evolutionary Biology of Parthenogenesis. 1128 Dordrecht: Springer Science+Business Media B.V. 1129 Schön I, Rossetti G, Martens K. 2009b. Darwinulid ostracods: ancient asexual scandals or 1130 scandalous gossip? In: Schon I, Martens K, van Dijk P, eds. Lost Sex: The Evolutionary 1131 Biology of Parthenogenesis: Dordrecht: Springer Science+Business Media B.V., 217- 1132 240. 1133 Schwander T. 2016. Evolution: The end of an ancient asexual scandal. Current Biology 26(6): 1134 R233–R235. 1135 Siepel H. 1994. Life-history tactics of soil microarthropods. Biology and Fertility of Soils 18(4): 1136 263-278. 1137 Siepel H. 1996. Biodiversity of soil microarthropods: the filtering of species. Biodiversity & 1138 Conservation 5(2): 251-260. 1139 Smith J. 1980. Selection for recombination in a polygenic model. Genetical Research 35(03): 1140 269-277. 1141 Speijer D, Lukes J, Elias M. 2015. Sex is a ubiquitous, ancient, and inherent attribute of 1142 eukaryotic life. Proceedings of the National Academy of Sciences of the United States of 1143 America 112(29): 8827-8834. 1144 Sterrer W. 2002. On the origin of sex as vaccination. Journal of Theoretical Biology 216(4): 387- 1145 396. 1146 Suttle CS. 2005. Viruses in the sea. Nature 437(7057): 356-361. 1147 Suttle CS. 2007. Marine viruses—major players in the global ecosystem. Nature Reviews 1148 Microbiology 5(10): 801-812.

1149 Suttle CS, Chan AM, Cottrell MT. 1990. Infection of phytoplankton by viruses and reduction of 1150 primary productivity. Nature 347(6292): 467-469. 1151 Thornhill R, Fincher C. 2013. The comparative method in cross-cultural and cross-species 1152 research. Evolutionary Biology 40(4): 480-493. 1153 Tobler M, Schlupp I, de Leon F, Glaubrecht M, Plath M. 2007. Extreme habitats as refuge from 1154 parasite infections? Evidence from an extremophile fish. Acta Oecologica 31(3): 270- 1155 275. 1156 Tokeshi M. 1999. Species Coexistence: Ecological and Evolutionary Perspectives. Oxford, UK: 1157 Blackwell Science Ltd. 1158 Turgeon J, Hebert P. 1994. Evolutionary interactions between sexual and all-female taxa of 1159 Cyprinotus (Ostracoda: Cyprididae). Evolution 48(6): 1855-1865. 1160 Van Dijk P. 2009. Apomixis: basics for non-botanists. In: Schön I, Martens K, van Dijk P, eds. 1161 Lost Sex: The Evolutionary Biology of Parthenogenesis. Dordrecht: Springer 1162 Science+Business Media B.V., 47-62. 1163 van Raay T, Crease T. 1995. Mitochondrial DNA diversity in an apomictic Daphnia complex 1164 from the Canadian High Arctic. Molecular Ecology 4(2): 149-161. 1165 Vanhaecke P, Siddall S, Sorgeloos P. 1984. International study on Artemia. XXXII. Combined 1166 effects of temperature and salinity on the survival of Artemia of various geographical 1167 origin. Journal of Experimental Marine Biology and Ecology 80(3): 259-275. 1168 von Uexküll J. 1982. The theory of Meaning. Semiotica 42(1): 25-82. 1169 Vrijenhoek R. 1984. The evolution of clonal diversity in Poeciliopsis. In: Turner B, ed. 1170 Evolutionary Genetics of Fishes. New York: Plenum Press, 399-429. 1171 Vrijenhoek R, Parker E. 2009. Geographical parthenogenesis: general purpose genotypes and 1172 frozen niche variation. In: Schön I, Martens K, van Dijk P, eds. Lost Sex: The 1173 Evolutionary Biology of Parthenogenesis: Dordrecht: Springer Science+Business Media 1174 B.V., 99-131. 1175 Walter D. 2009. Suborder Endeostigmata. In: Krantz G, Walter D, eds. A manual of acarology. 1176 Lubbock: Texas Tech University Press, 421-429. 1177 Walter D, Lindquist E, Smith I, Cook D, Krantz G. 2009. Order Trombidiformes. In: Krantz G, 1178 Walter D, eds. A manual of acarology. Lubbock: Texas Tech University Press, 233-420.

1179 Wallwork, JA. 1970. Ecology of soil animals. London, UK: McGRAW-HILL Publishing 1180 Company Limited. 1181 West S, Lively C, Read A. 1999. A pluralist approach to sex and recombination. Journal of 1182 Evolutionary Biology 12(6): 1003-1012. 1183 Williams GC. 1975. Sex and evolution. Princeton, NJ: Princeton University Press. 1184 Wilson CG. 2011. Desiccation-tolerance in bdelloid rotifers facilitates spatiotemporal escape 1185 from multiple species of parasitic fungi. Biological Journal of the Linnean Society 1186 104(3): 564-574 1187 Wommack KE, Colwell RR. 2000. Virioplankton: Viruses in Aquatic Ecosystems. Microbiology 1188 and Molecular Biology Reviews 64(1): 69-114.